Drug addiction is a chronic, devastating, but treatable disorder, for which there exists a growing armamentarium of evidence-based interventions, including pharmacotherapies and psychotherapies. A core principle of drug addiction treatment, however, states that no single treatment is appropriate for everyone; rather, treatments need to be adjusted based on patient characteristics and response in order to be maximally effective. Ideally, clinicians would identify a sequence of interventions that works best across different stages of addiction treatment, from abstinence initiation to relapse prevention. Adaptive treatment interventions have been used successfully to inform this sequential clinical decision-making process. For cocaine use disorders (CUD), the most potent intervention currently available for initiating abstinence is behavior therapy using contingency management (CM) procedures. Intensive CM has been shown to produce initial cocaine abstinence rates of 40%, unmatched by all other forms of behavioral or pharmacological treatment, making it a prototypical first-line therapy for CUD. Importantly, achievement of initial abstinence predicts future abstinence. For the clinician, these research findings translate into a straightforward question: Can we drive CM response rates even higher with targeted adjunctive interventions? The proposed sequential, multiple assignment, randomized trial (SMART) will provide the data needed to answer this question. First, we will determine whether Acceptance and Commitment Therapy (ACT) in combination with CM increases initial treatment response rates. We hypothesize that four weeks of treatment with ACT+CM will produce higher abstinence rates than initial treatment combining standard Drug Counseling with CM (DC+CM). The hypothesized synergism of ACT+CM on primary treatment mechanisms of experiential avoidance and reward sensitivity, respectively, will be examined. Second, for patients who do not respond to initial treatment, we will examine whether dopamine-targeted pharmacotherapy is an effective augmentation strategy. Specifically, we hypothesize that continued ACT+CM treatment with levodopa-carbidopa augmentation will be most effective in promoting abstinence relative to treatment combinations involving continued DC and/or placebo. Third, for patients who respond to initial treatment, we will assess the relative benefit of continued treatment with ACT+CM, as compared to DC+CM, to prevent relapse. ACT emphasizes goal-directed actions based on values that are intrinsically motivating, and is thereby expected to be a more effective intervention for extending the duration of abstinence following initial treatment with intensive CM. In summary, results from this Stage II (PA-13-078) project should support a Go/No Go decision about further Phase III, confirmatory studies. The primary aims of this project address an important NIDA research priority and have the potential to significantly impact how we tailor treatment of CUD to maximize outcomes.